Embodiments of the present invention relate to request scheduling and, more specifically, to scheduling requests in an execution environment.
In many cases, a middleware environment, such as WebSphere, hosting a software solution is configured to handle a certain load of requests. For instance, the various resources for the environment might be selected based on expected capacity, so as to achieve a desired level of performance. Despite this, the environment can suffer capacity problems, such as input or output bottlenecks or high memory of processor utilization. Such circumstances may have various causes, such as, for example, a sudden increase in requests, a network or disk failure, or a slowdown in a backend system.
The result of such capacity problems is that service-level agreements may be violated. For instance, increases in request processing times may occur, in violation of agreements with customers issuing those requests.